A Systems Genetics Approach to Dyslipidemia in Children and Adolescents

Abstract

Elevated triglycerides (TG) or low high density lipoprotein cholesterol (HDL-C) levels are common cardiometabolic risk factors in children. From a systems genetics standpoint, Visualization of Statistical Epistasis Networks (ViSEN) is a nonparametric entropy-based method that can characterize the global structure of interacting genetic factors. We identified a novel set of connected genetic and cardiometabolic risk factors with strong and significant interaction effects on two important dyslipidemia phenotypes (low HDL-C and high TG) in children and adolescents. Study participants were recruited from five schools in Istanbul, Turkey (n=360; 170 boys, 190 girls). Participants with TG levels≥75th and HDL-C levels≤25th percentile were defined as 'high TG' and 'low HDL-C', respectively. We genotyped participants for six single nucleotide polymorphisms (SNPs) in five genes with known associations to lipid levels (rs328 in LPL, rs708272 in CETP, rs1800588 in LIPC, rs1800977 in ABCA1, rs1799941 and rs6257 in SHBG gene). ViSEN was used to identify associations with dyslipidemia phenotypes. There were 71 (50 males, 21 females) and 93 (60 males and 33 females) subjects with low HDL-C and high TG, respectively. Biological variables including age, gender, and BMI were significantly associated with both phenotypes (p<0.001). Importantly, a single SNP, rs708272, was associated with low HDL-C (IG=2.24%, p=0.026). Pairwise and higher order interaction analyses in the full dataset for low HDL-C and high TG revealed the largest effects in the models containing rs1800977, rs708272, age (IG=6.20%, p=0.046) and rs1800588, age, BMI (IG, 3.06%, p=0.022), respectively. In conclusion, the present study brings us a step closer to a systems genetic approach in understanding lipid phenotypes in children. Further efforts can integrate population and laboratory-based studies, hence accelerate the preventive medicine efforts.

Pairwise and Higher-Order Epistatic Interactions Have a Significant Impact on Bronchodilator Drug Response in African American Youth with Asthma

“…Research suggests that ViSEN is more powerful than standard regression-based methods for detecting nonlinear, non-additive interaction effects. 19,26,33 Effects of single explanatory variables, which ViSEN defines as main effects, on phenotype status is calculated using Mutual Information (MI). MI, derived from information theory, quantifies the reduction in uncertainty about the distribution of one variable given an understanding of the other.…”

An epistatic interaction between pre-natal smoke exposure and socioeconomic status has a significant impact on bronchodilator drug response in African American youth with asthma

“…ViSEN is a statistical program that implements the information theory metrics of MI and IG to perform network-based analyses that quantify and visualize pairwise and higherorder epistatic interactions in case control studies. Research suggests that ViSEN is more powerful than standard regression-based methods for detecting additive non-additive interaction effects [19,26,33]. ViSEN measures the effects of single explanatory variables on case/control status in terms of MI, and the strength of pairwise interaction effects on case/control status is quantified using IG.…”

“…ViSEN is also able to identify and quantify both additive and non-additive interactions and provide intuitive visualization of the potentially complex relationships between large numbers of variables using a network-based approach. Additionally, ViSEN has been shown to be more powerful than standard regression-based methods in detecting interaction effects, suggesting that ViSEN is a promising investigative tool that can be applied to studies of complex traits such as BDR [20,[24][25][26][27][28].…”

Polymorphisms of lipid metabolism enzyme-coding genes in patients with diabetic dyslipidemia

“…Contrary to the findings of Yılmaz et al (27), we did not find any association between the CETP rs708272 polymorphism and biochemical parameters such as triglyceride or HDL cholesterol levels in our diabetic dyslipidemia patient group. A recent study demonstrated that a CETP rs708272 gene variant is associated with low HDL cholesterol phenotype in Turkish children (28, 29). As our patient population was on lipid-lowering drugs, insulin, and oral antidiabetics, the effects of polymorphisms on lipid parameters and glucose might be masked by drug treatments.…”